Uninformative prior¶

In case if we do not have a strong belief about our prior, we should let the data speak for itself. This is a prior that has only weak influence on the data.

Unfortunatelly coosing a uninformative prior can be hard, but there is a general purpose technique, whichs result is called the Jeffreys prior.

In general it is a good practice to perform sensitivity analysis. This is to check how sensitive our model is to change of prior.